Google LLC announced some updates to its cloud platform today that should lead to a significant reduction in network latency for its customers and faster, more secure connections for those seeking higher availability.

The first update relates to part of the core infrastructure of Google’s cloud. The internet giant said it’s updating the Andromeda software-defined network virtualization stack that underpins its Google Cloud Platform to version 2.1, bringing about a 40 percent reduction in latency between Compute Engine virtual machines.

Launched in 2014, Andromeda serves as the orchestration point in Google’s cloud for provisioning, configuring and managing virtual networks. It’s used to deliver networking services with high performance, availability, isolation, and security. For example, cloud platform firewalls, routing, and forwarding rules all leverage the underlying internal Andromeda APIs and infrastructure.

In a blog post, Google software engineer Jake Adriaens said the new Andromeda release means the company has managed to reduce network latency for Compute Engine VMs by a factor of eight since the stack was first released.

“This kind of network performance is especially important as more applications move into the cloud and are accessed via web browsers,” Adriaens wrote. “While the headline metric is often bandwidth, network latency is frequently the more important determiner of application performance.”

He added that lower latency is a key requirement for tasks such as financial transactions, ad tech, video, gaming, high-performance computing and in-memory databases.

Adriaens said Andromeda’s new lower latencies were achieved by effectively bypassing the hypervisor, which is a part of the stack that abstracts operating systems and applications away from the underlying hardware so virtual machines can share system resources such as processor cycles, memory space and network bandwith. Previously the hypervisor was used as a bridge between Andromeda and guest VMs. With the update, however, guest VMs and Andromeda now can communicate directly.

In addition to Andromeda, Google this week also announced updates and general availability for its Cloud Dedicated Interconnect service. Unveiled in September, Cloud Dedicated Interconnect enables customers to connect to its data centers via enterprise-grade connections with higher availability and lower latency than their existing Internet connections.

Google Product Manager John Veizades said in a second blog post that Cloud Dedicated Interconnect is now ready for production-grade workloads, covered by a service level agreement. In addition, the service has been updated to support Cloud Router Global Routing, a new feature that allows subnets in GCP to be accessible from any on-premises network through the Google network.

Veizades said Cloud Dedicated Interconnect is now being offered from four new locations, including Atlanta, Mumbai, Munich and Montreal, in addition to the existing locations detailed on the product page. The service is priced at $1,700 a month per 10 Gbps link.

Image: RettungsgasseJETZTde/pixabay